The Missing Line Item in Loneliness Research: Cost

Cigna’s 2020 workplace report is one of the most quoted statistics in loneliness journalism: lonely workers miss work twice as often due to illness and five times as often due to stress. It is a striking number, and it is not a cost figure. It is a self-reported frequency ratio from a cross-sectional survey. No dollar amount attaches to it in the source. Yet the number routinely gets converted, in secondary coverage, into implied productivity losses — a conversion the underlying data cannot support, because the survey never measured actual absence days, employer costs, or a comparison group tracked over time.

This is the pattern across the loneliness literature generally, and it becomes more acute wherever a technology-based intervention is proposed as the fix: an app, a platform, a matching algorithm. The health evidence for social disconnection as a risk factor is comparatively strong. The economic evidence for what it costs, and for what any given intervention saves, is thin almost everywhere it has been asked for.

What the health evidence establishes, and what it doesn’t

Start with what is solid. Julianne Holt-Lunstad’s meta-analyses put social isolation and loneliness in the same statistical territory as recognized mortality risk factors, and the American Heart Association’s 2022 scientific statement quantified the association with cardiovascular outcomes specifically: roughly 29% increased risk of heart attack and death from heart disease, 32% increased risk of stroke. These are associational findings from large pooled samples, not causal proof, and the AHA statement is unusually candid about this. It names the absence of intervention evidence as the central gap in the field — not a footnote, the headline gap. That single sentence in a document from a major medical association is a more useful guide to the state of the science than most of what gets built on top of it.

Holt-Lunstad’s own 2021 review argues that social connection should sit alongside diet, exercise, and smoking cessation in preventive health frameworks. That is a reasonable position given the mortality data. But diet and smoking cessation both have decades of health-economic modeling behind them — cost per quality-adjusted life year, cost offsets from reduced hospitalization, comparative studies of intervention formats. Social connection has almost none of that. The 2023 Surgeon General’s advisory, which did more than any single U.S. document to establish loneliness as a public health priority, compares the mortality risk of disconnection to smoking up to 15 cigarettes a day. It does not, and does not claim to, offer a comparable economic accounting.

Social prescribing: the closest thing to cost data, and its limits

If there is a place to look for economic evidence on connection interventions, it is social prescribing — the UK-originated practice of a clinician referring a patient to a community activity, group, or service rather than, or alongside, medical treatment. It has been running long enough, and is embedded in enough health systems, that someone has tried to count the savings.

A 2021 systematic review in Perspectives in Public Health found that of nine included studies, three reported reductions in use of GP appointments, emergency care, social worker time, or inpatient stays. That is a real signal, and it is the kind of proxy that health economists use to build cost arguments. But three out of nine is not a consistent finding, and “reported a reduction” in a small, heterogeneous set of studies is a long way from a costed comparison against usual care, with confidence intervals, over a defined time horizon. A parallel review the same year, covering wellbeing outcomes more broadly, reports gains in self-esteem and confidence but explicitly flags limited trial evidence and wide heterogeneity across programs as a constraint on what can be concluded. A later qualitative synthesis, from 2022, adds that participants describe benefit from structured, purposeful activity rather than social contact alone — useful for program design, but it is qualitative testimony, not a cost figure.

Put together, social prescribing has more economic evidence attached to it than almost any other connection intervention, and what exists still falls well short of a defensible cost-effectiveness case. That is the high-water mark, not the exception.

What the randomized evidence looks like when it exists at all

Most of the loneliness intervention literature is uncontrolled: pre-post surveys of people who opted into a program, with no comparison group and no way to separate the effect of the activity from the effect of being the kind of person who signs up for it. The 2024 HEAL-HOA trial, published in The Lancet Healthy Longevity, is one of the few genuine randomized controlled trials in this space — testing prosocial engagement and volunteering against a control condition among lonely older adults in Hong Kong. Its value is structural: it is a real trial in a field mostly made up of program evaluations. But it was designed to test whether the intervention reduced loneliness, not what it cost to deliver relative to any offset in downstream health service use. An RCT that measures the primary outcome and not the economic one still leaves the cost question exactly where it was.

Where the technology platforms sit

This is where the gap becomes most visible for anyone assessing technology as a solution rather than health policy as one. Meetup, the longest-running mainstream platform for recurring interest-based groups, is commonly cited by member count — around 60 million — which is a usage figure, not an outcome or a cost figure. Timeleft, which organizes weekly dinners matching strangers by personality and reports having seated more than 3 million guests, likewise reports scale rather than measured effect on loneliness or any downstream health or productivity outcome. Buffet, covered by TechCrunch in 2024 as a Los Angeles-launched app that used an interest and personality questionnaire to match a person with another person and a place to meet, is earlier-stage still; the reporting describes the product and its freemium model, not any independent measurement of whether it reduces loneliness, let alone what that would save a health system or an employer.

None of this is a criticism specific to these platforms. It reflects the fact that essentially no consumer social-connection technology has been subjected to the kind of trial design, follow-up period, and economic evaluation that would let anyone say what a reduction in loneliness through the platform is worth in health-system or workplace terms. Gallup’s 2024 workplace data shows fully remote employees reporting loneliness at 25% against 16% for fully on-site staff — a real and policy-relevant gradient — but it is a snapshot comparison across work arrangements, not a study of whether any specific connection tool changes that gradient or at what cost.

Why this matters for how these claims get used

The National Academies’ 2020 report calling on health systems to routinely screen older adults for isolation is built on solid prevalence and mortality evidence — roughly a quarter of adults 65 and older are considered isolated, and the mortality link is well established. But a screening mandate creates a downstream question the mortality data cannot answer: what does the intervention that follows a positive screen cost, and does it produce savings that justify the screening infrastructure. That is an economic question, and right now the field mostly does not have an economic answer for it, technology-based or otherwise.

A study that would move this forward would need three things current work lacks nearly across the board: a randomized design with a genuine control arm, a monetized cost of delivering the intervention including any technology or staffing overhead, and a tracked economic outcome over a defined follow-up period — health service utilization, employer absenteeism records rather than self-report, or both. Until something like that exists for a specific platform or program, the honest description of the economic case for any connection intervention, technological or otherwise, is that it has not yet been made.